mh-openai-whisper-api
от mohdalhashemi98-hue
AI & LLMsКратко
Transcribe audio via OpenAI Audio Transcriptions API (Whisper).
Описание
Что делает этот навык
Transcribes audio files by sending them to OpenAI's Whisper API endpoint. Supports multiple audio formats, language hints, speaker prompts, and JSON or plain text output.
Uses OpenAI's hosted Whisper model via a single curl call, so no local model setup or GPU is needed.
Когда его использовать
- Transcribing recorded meetings or interviews
- Converting voice memos to text
- Generating subtitles from audio clips
- Transcribing podcast episodes
- Extracting text from language learning recordings
Пример сценария
User asks: transcribe my interview recording at /recordings/interview.m4a with speaker names Peter and Daniel
- 1
Receives audio file path and speaker context
- 2
Calls transcribe.sh with --prompt flag set to 'Speaker names: Peter, Daniel'
- 3
Sends multipart form request to /v1/audio/transcriptions with whisper-1 model
- 4
Writes transcript to interview.m4a.txt in the same directory
Plain text transcript saved to /recordings/interview.m4a.txt
Требования
- OPENAI_API_KEY environment variable
Установка
clawhub install mohdalhashemi98-hue/mh-openai-whisper-apiВыполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.